Abstract

This research analyzes the main factors that lead to the occurrence of traffic conflicts on the urban highway Autopista Central in Santiago, Chile. We propose a vehicle-by-vehicle approach in which we estimate the probability that each vehicle passing through each gate of Autopista Central has a conflict. This approach allows us to study variables at an individual level that have not been analyzed before using full real-world data, such as driver (age and sex) and vehicle information (type and age). We propose a Pseudo Time-to-Collision (PTTC) as a surrogate safety measure, defining different critical PTTC thresholds for each gate. Subsequently, a logistic regression model is built to understand the input variables' influence on the probability of conflict occurrence. Our results show that men and young drivers are more likely to have a conflict. In addition, the age of the car is negatively correlated with the occurrence of conflicts.